Transmission lines is an important part of the grid, the primary means of normal operation of the power grid to ensure periodic inspection of transmission line. With the development of UAV technology, aerial drones being used more and more transmission line inspection work. This article is an operating unit of existing transmission lines drone shot images, research on high-voltage transmission lines for the technical inspection for the realization of the use of drones for automatic inspection provides an important technical foundation. Transmission Line state detection based on analysis of bionic vision has become the most important issue currently "strong smart grid", to enhance the existing grid security is of great significance. Conduct based on the identification and monitoring method UAV aerial transmission line inspection image depth research to patrol the image transmission line inspection process collected for the study, the use of intelligent analysis, pattern recognition, computer vision soft measurement technology, transmission line equipment to achieve recognition and fuzzy abnormal state discrimination. The main methods include:inspection of the video image preprocessing, key frame extraction sequence, fault characteristics under complex natural background image data transmission lines massive precision intelligent processing, transmission lines and key components of extraction, automatic fault identification and transmission equipment stateful inspection and so on.The main research work is as follows:1, according to the characteristics of the aerial image, image pre-study method using gray image, brightness balance, contrast enhancement, image denoising, image deblurring techniques, thereby improving the quality aerial imagery.2, for aerial image transmission line characteristics of the study based on Canny edge detection and LSD straight line detection method for detecting transmission technology, to achieve the automatic extraction of the transmission line.3, in order to solve the intrusion transmission lines around the study of detection and tracking moving targets to achieve a frame difference, background modeling moving object detection method based on Mean Shift and Particle Filter Method moving target tracking technology.The proposed method, the image is analyzed daily accumulation of tests to prove its good results. |